dennybritz / rnn-tutorial-rnnlm

Recurrent Neural Network Tutorial, Part 2 - Implementing a RNN in Python and Theano
Apache License 2.0
895 stars 468 forks source link

Why the timestep is not used in the forward_propagation #24

Open Tongjilibo opened 5 years ago

Tongjilibo commented 5 years ago

When i use lstm in keras or tensorflow, they have different data strctures. When i compared this tutorial with keras, i found the 'timestep' is only used in the backward_propagation, but not in the forward_propagation is this right?